Abstract: The existing typical link evaluation measures in wireless sensor networks are investigated, since the evaluation results of RSSI and LQI are inaccurate if they have just been evaluated for directly once. A new evaluation approach is therefore proposed taking the mean RSSI and mean LQI as the measures in link evaluation so as to avoid such as inaccurate evaluation result. Taking the nodes of MicaZ sensors as an experimental platform to investigate the communication links in wireless sensor network, a link evaluation model based on Gauss distribution is proposed and verified with lots of sampled values analyzed statistically. The modeling approach suggested has the generality that the Gauss link evaluation model thus developed will provide reference for not only applications, but also the simulation investigation on upper layer protocols in wireless sensor networks.